压力容器2025，Vol.42Issue(11)：75-86,12.DOI:10.3969/j.issn.1001-4837.2025.11.009
基于声发射特征的复合材料层合板多元统计分析与聚类评价
Multivariate statistical analysis and clustering evaluation of composite laminates based on acoustic emission features
摘要
Abstract
To meet the needs for composite structure integrity assessment and identification of different damage mechanisms,a method based on acoustic emission(AE)features and unsupervised clustering is proposed.Using acoustic emission data from glass fiber laminates with different interface fiber orientations under double cantilever beam tests,multivariate statistical methods are employed to select AE features containing the most damage information.The clustering by fast search and find of density peaks(CFSFDP)algorithm,a density-based clustering approach,is applied to cluster the data after dimensionality reduction by principal component analysis,effectively identifying three damage mechanisms in composite laminates with varying interface fiber orientations.The results show that through analysis of factors affecting the stability of the clustering algorithm,the choice of distance metric has little impact on the clustering results,while the cut-off parameter in the range of 1%～2%enables the algorithm to identify the acoustic emission signals of each damage mechanism with good accuracy.The proposed method provides a reference for damage mechanism identification in composite materials.关键词
复合材料层合板/声发射/损伤模式识别/损伤表征/多元统计分析Key words
composite laminate/acoustic emission/damage mode identification/damage characterization/multivariate statistical analysis分类
机械制造引用本文复制引用
杨畅,孙博文,李伟,蒋鹏,左坤平..基于声发射特征的复合材料层合板多元统计分析与聚类评价[J].压力容器,2025,42(11):75-86,12.基金项目
黑龙江省自然科学基金(PL2025E013) （PL2025E013）
